feat: add --scrape.file flag to scrape metrics from local files

README.md

@@ -24,6 +24,7 @@ for creating an intuitive and interactive command-line interface.
 - [x] Allow to select a specific metric to inspect, and show its series.
 - [x] Metric search with fuzzy search.
 - [x] [HTTP configuration file](https://prometheus.io/docs/prometheus/latest/configuration/configuration/#http_config) to support custom HTTP client options like basic auth, custom headers, proxy configs, etc.
+- [x] Support reading metrics from local files, to help analyze metrics that have already been collected.
 
 ## Planned Features
 

cmd/cardinality.go

@@ -397,9 +397,19 @@ func registerCardinalityCommand(app *extkingpin.App) {
 		_ bool,
 	) error {
 		scrapeURL := opts.ScrapeURL
+		scrapeFile := opts.ScrapeFile
 		timeoutDuration := opts.Timeout
 		httpConfigFile := opts.HttpConfigFile
 
+		if scrapeURL == "" && scrapeFile == "" {
+			return errors.New("No URL or file provided to scrape metrics. " +
+				"Please supply a target to scrape via `--scrape.url` or `--scrape.file` flags.")
+		}
+
+		if scrapeURL != "" && scrapeFile != "" {
+			return errors.New("The flags `--scrape.url` and `--scrape.file` are mutually exclusive.")
+		}
+
 		metricTable := newModel(nil, opts.OutputHeight, logger)
 		p := tea.NewProgram(metricTable)
 		metricTable.program = p
@@ -424,7 +434,8 @@ func registerCardinalityCommand(app *extkingpin.App) {
 
 			level.Info(logger).Log(
 				"msg", "scraping",
-				"url", scrapeURL,
+				"scrape_url", scrapeURL,
+				"scrape_file", scrapeFile,
 				"timeout", timeoutDuration,
 				"max_size", maxSize,
 				"http_config_file", httpConfigFile,
@@ -433,6 +444,7 @@ func registerCardinalityCommand(app *extkingpin.App) {
 			t0 := time.Now()
 			scraper := scrape.NewPromScraper(
 				scrapeURL,
+				scrapeFile,
 				logger,
 				scrape.WithTimeout(timeoutDuration),
 				scrape.WithMaxBodySize(maxSize),

cmd/options.go

@@ -10,6 +10,7 @@ import (
 
 type Options struct {
 	ScrapeURL      string
+	ScrapeFile     string
 	OutputHeight   int
 	MaxScrapeSize  string
 	Timeout        time.Duration
@@ -25,10 +26,14 @@ func (o *Options) MaxScrapeSizeBytes() (int64, error) {
 }
 
 func (o *Options) AddFlags(app extkingpin.AppClause) {
-	app.Flag("scrape-url", "URL to scrape metrics from").
-		Required().
+	app.Flag("scrape.url", "URL to scrape metrics from").
+		Default("").
 		StringVar(&o.ScrapeURL)
 
+	app.Flag("scrape.file", "File to scrape metrics from").
+		Default("").
+		StringVar(&o.ScrapeFile)
+
 	app.Flag("timeout", "Timeout for the scrape request").
 		Default("10s").
 		DurationVar(&o.Timeout)

pkg/scrape/scraper.go

@@ -5,6 +5,7 @@ import (
 	"fmt"
 	"io"
 	"net/http"
+	"os"
 	"regexp"
 	"strconv"
 	"strings"
@@ -23,6 +24,7 @@ import (
 type PromScraper struct {
 	httpConfigFile        string
 	scrapeURL             string
+	scrapeFilePath        string
 	timeout               time.Duration
 	logger                log.Logger
 	series                map[string]SeriesSet
@@ -56,7 +58,7 @@ func WithHttpConfigFile(file string) ScraperOption {
 	}
 }
 
-func NewPromScraper(scrapeURL string, logger log.Logger, opts ...ScraperOption) *PromScraper {
+func NewPromScraper(scrapeURL string, scrapeFile string, logger log.Logger, opts ...ScraperOption) *PromScraper {
 	scOpts := &scrapeOpts{
 		timeout:        10 * time.Second,
 		maxBodySize:    10 * 1024 * 1024,
@@ -69,6 +71,7 @@ func NewPromScraper(scrapeURL string, logger log.Logger, opts ...ScraperOption)
 
 	return &PromScraper{
 		scrapeURL:      scrapeURL,
+		scrapeFilePath: scrapeFile,
 		logger:         logger,
 		timeout:        scOpts.timeout,
 		maxBodySize:    scOpts.maxBodySize,
@@ -79,6 +82,58 @@ func NewPromScraper(scrapeURL string, logger log.Logger, opts ...ScraperOption)
 }
 
 func (ps *PromScraper) Scrape() (*Result, error) {
+	if ps.scrapeFilePath != "" {
+		return ps.scrapeFile()
+	}
+
+	return ps.scrapeHTTP()
+}
+
+func (ps *PromScraper) scrapeFile() (*Result, error) {
+	var (
+		seriesSet        map[string]SeriesSet
+		seriesScrapeText SeriesScrapeText
+	)
+
+	// Don't use os.ReadFile(); manually open the file so we can create an
+	// io.LimitReader from the file to enforce max body size.
+	f, err := os.Open(ps.scrapeFilePath)
+	if err != nil {
+		return &Result{}, fmt.Errorf("Failed to open file %s to scrape metrics: %w", ps.scrapeFilePath, err)
+	}
+	defer f.Close()
+
+	body, err := io.ReadAll(io.LimitReader(f, ps.maxBodySize))
+	if err != nil {
+		return &Result{}, fmt.Errorf("Failed reading file %s to scrape metrics: %w", ps.scrapeFilePath, err)
+	}
+
+	if int64(len(body)) >= ps.maxBodySize {
+		level.Warn(ps.logger).Log(
+			"msg", "metric file body size limit exceeded",
+			"limit_bytes", ps.maxBodySize,
+			"body_size", len(body),
+		)
+		return &Result{}, fmt.Errorf("metric file body size exceeded limit of %d bytes", ps.maxBodySize)
+	}
+
+	// assume that scraping metrics from a file implies they're in text format.
+	contentType := "text/plain"
+	ps.lastScrapeContentType = contentType
+	seriesSet, scrapeErr := ps.extractMetrics(body, contentType)
+	if scrapeErr != nil {
+		return &Result{}, fmt.Errorf("failed to extract metrics from file: %w", scrapeErr)
+	}
+	seriesScrapeText = ps.extractMetricSeriesText(body)
+
+	return &Result{
+		Series:           seriesSet,
+		UsedContentType:  contentType,
+		SeriesScrapeText: seriesScrapeText,
+	}, nil
+}
+
+func (ps *PromScraper) scrapeHTTP() (*Result, error) {
 	var (
 		seriesSet        map[string]SeriesSet
 		scrapeErr        error